Step #2: Deploy Sample Applications

As the NVIDIA LaunchPad system is already configured with NVIDIA Cloud Native Stack, you can run sample applications with GPUs on it. For more information, please refer to the Validate Sample Application with NVIDIA Cloud Native Stack.

Run the below command on the NVIDIA LaunchPad system to verify the nvidia-smi on Kubernetes.

            

            
cat <<EOF | tee nvidia-smi.yaml
apiVersion: v1
kind: Pod
metadata:
name: nvidia-smi
spec:
restartPolicy: OnFailure
containers:
- name: nvidia-smi
image: "nvidia/cuda:11.8.0-base-ubuntu20.04"
args: ["nvidia-smi"]
EOF
        

1. Execute the below command to create a nvidia-smi pod

   Copy

   Copied!

               
   
               
   Kubectl apply -f nvidia-smi.yaml
           

   
   

2. execute the below command to see the result.

   Copy

   Copied!

               
   
               
   kubectl logs pod/nvidia-smi
           

   
   

Expected Output:

            

            
Wed Oct 27 17:17:10 2022
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 520.61.07    Driver Version: 520.61.07    CUDA Version: 11.8     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            On   | 00000000:14:00.0 Off |                  Off |
| N/A   47C    P8    16W /  70W |      0MiB / 16127MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
        

1. Create a cuda-sampled.yaml with the help of the below command.

   Copy

   Copied!

               
   
               
   cat <<EOF | tee cuda-samples.yaml
   apiVersion: v1
   kind: Pod
   metadata:
   name: cuda-vector-add
   spec:
   restartPolicy: OnFailure
   containers:
   - name: cuda-vector-add
   image: "k8s.gcr.io/cuda-vector-add:v0.1"
   EOF
           

   
   

2. Execute the below command to create a sample pod

   Copy

   Copied!

               
   
               
   Kubectl apply -f cuda-samples.yaml
           

   
   

3. Once the container run completes, execute the below command to see the result.

   Copy

   Copied!

               
   
               
   kubectl logs pod/cuda-vector-add